Multi-Objective Parameter Selection for Classifiers

Christoph Müssel, Ludwig Lausser, Markus Maucher and Hans A. Kestler

Journal of Statistical Software, 2012, vol. 046, issue i05

Abstract: Setting the free parameters of classifiers to different values can have a profound impact on their performance. For some methods, specialized tuning algorithms have been developed. These approaches mostly tune parameters according to a single criterion, such as the cross-validation error. However, it is sometimes desirable to obtain parameter values that optimize several concurrent - often conflicting - criteria. The TunePareto package provides a general and highly customizable framework to select optimal parameters for classifiers according to multiple objectives. Several strategies for sampling and optimizing parameters are supplied. The algorithm determines a set of Pareto-optimal parameter configurations and leaves the ultimate decision on the weighting of objectives to the researcher. Decision support is provided by novel visualization techniques.

Date: 2012-01-30
References: View complete reference list from CitEc
Citations: View citations in EconPapers (3)

Downloads: (external link)
https://www.jstatsoft.org/index.php/jss/article/view/v046i05/v46i05.pdf
https://www.jstatsoft.org/index.php/jss/article/do ... unePareto_2.2.tar.gz
https://www.jstatsoft.org/index.php/jss/article/do ... ile/v046i05/v46i05.R
https://www.jstatsoft.org/index.php/jss/article/do ... 6i05/parkinsons.data

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:jss:jstsof:v:046:i05

DOI: 10.18637/jss.v046.i05

Access Statistics for this article

Journal of Statistical Software is currently edited by Bettina Grün, Edzer Pebesma and Achim Zeileis

More articles in Journal of Statistical Software from Foundation for Open Access Statistics
Bibliographic data for series maintained by Christopher F. Baum ().